Read just about any news report on swine flu deaths, and you'll come across a line that claims "36,000 people die each year from flu-related causes." It sounds authoritative. It's even a nice, round number. But where is this number coming from? And is it based on any actual science?

This statistic is being paraded around by almost everybody, as if to say that swine flu isn't so bad because regular flu kills so many people each year anyway. The truth is that the only standard by which the CDC and WHO are quoting deaths from swine flu is if they are confirmed deaths from a particular viral strain. To them, if a death has not been confirmed in their labs, it does not count as a death from that flu.

Got that? Only "confirmed" deaths count. And they must be confirmed in a laboratory using a rigorous method of comparing samples taken from the deceased with a known database of viral patterns.

As it turns out, virtually none of the 36,000 people said to die from regular flu each year have been confirmed in any lab whatsoever.

Thus, according to the guidelines of the CDC and WHO, they don't count. Based on their own rules, it is technically accurate to say that regular flu kills virtually no one. It's not true, of course, because people do die from the "regular flu" each year, but it is technically accurate according to the CDC and WHO rules for scientific evidence.

Again, that's because nearly all of these "regular flu" deaths aren't confirmed by a CDC or WHO-recognized lab. Thus, they have no scientific standing.

Infectious disease double standard

I find it interesting that when talking about swine flu, the criteria for inclusion in statistics is positive identification in a rigorous laboratory. But when talking about regular flu, the criteria for inclusion is -- technically speaking -- anybody's wild guess.

The 36,000 number, it turns out, was pulled out of thin air. It has no scientific validity whatsoever, even according to the CDC's own standards.

I tracked down the origins of this number on CDC.gov, by the way. Turns out it was an estimate derived by the CDC in 2003 (http://www.cdc.gov/od/oc/media/pres...).

It's an estimate, mind you, not a "confirmed" number of deaths. And that estimate has stayed exactly the same through 2003, 2004, 2005, 2006, 2007, 2008 and 2009. Not a budge. Before the number was 36,000, it was 20,000 for many years. That tells you right off the bat this isn't some confirmed laboratory number -- it's a guesstimate!

I'm not disagreeing with the number. It's probably a fairly accurate guess (the CDC folks are a smart bunch). But it doesn't meet the criteria by which these infectious disease organizations report influenza deaths.

As the CDC even says on their own website, "This estimate came from a 2003 study published in the Journal of the American Medication Association (JAMA), which looked at the 1990-91 through the 1998-99 flu seasons [10]. Statistical modeling was used to estimate how many flu-related deaths occurred among people whose underlying cause of death on their death certificate was listed as a respiratory or circulatory disease. During these years, the number of estimated deaths ranged from 17,000 to 52,000."

In other words, they took a look at how many people died from respiratory or circulatory disease, and from that they extrapolated "flu-related deaths."